This invention relates to a motor control device and an image forming system.
Conventionally, a control system is known in which disturbance is estimated by a disturbance observer and a control input is corrected based on the estimated disturbance. FIG. 8 is a block diagram of a conventional motor control system 900 using a disturbance observer 910.
The control system shown in FIG. 8 is designed as follows. First, a transfer function G(s) of a controlled object (plant) is derived. Then, on the assumption that a relationship between a control output y and a control input u satisfies y=G(s)·u, a controller 920 is designed to output the control input u corresponding to a target control output yr.
The control input u outputted from the controller 920 is corrected by adding thereto a correction amount d* generated in the disturbance observer 910. Disturbance is suppressed by inputting a corrected control input uc=u+d* into the controlled object, and the control output y is controlled to the target control output yr.
Specifically, the disturbance observer 910 generates a correction signal (uc−1/G(s)·y) for suppressing disturbance based on an inverse model 1/G(s) of the transfer function G(s) and the control input uc. The disturbance observer 910 filters the correction signal through a low-pass filter 930, and outputs a filtered correction signal as a correction signal representing the above correction amount d*.
A reason why the correction signal (uc−1/G(s)·y) is filtered through the low-pass filter 930 is that the transfer function G(s) of the controlled object exhibits such frequency characteristics that the higher a frequency of a high-frequency component in the correction signal is, the lower a gain becomes.
When the transfer function G(s) exhibits such frequency characteristics, it should be appreciated that the inverse model 1/G(s) exhibits such frequency characteristics that the higher a frequency of a high-frequency component in the correction signal is, the higher a gain becomes. On the other hand, a measurement signal of the control output y includes a lot of observation noise in a high-frequency component.
Accordingly, when the measurement signal of the control output y is inputted into the inverse model 1/G(s) to generate the correction signal (uc−1/G(s)·y), the observation noise is amplified in the correction signal. Therefore, the low-pass filter 930 is used to suppress influence of the observation noise and accurately achieve disturbance suppression.